Neuroscience increasingly requires the integration of sophisticated experimental and quantitative approaches. The development of increasingly sophisticated models and the analysis of massive sources of data routinely push neuroscientists to the limits of their quantitative and analytical abilities. In this way, computational neuroscience has become an element of the mainstream of neuroscience research. However, few training programs have adapted to this increase in the importance of quantitative approaches in neuroscience, leaving many students under prepared to exploit future opportunities in the field. Here we propose training programs involving faculty from more than ten departments at Carnegie Mellon University and the University of Pittsburgh that seek to move computational neuroscience into the mainstream of the field. Specifically, we seek to: 1) expose hundreds of undergraduate students in biomedical fields and hundreds of students in quantitative disciplines to computational neuroscience each year. 2) develop a comprehensive research and education program that provide excellent in depth training in computational neuroscience to 10-15 undergraduates from a variety of disciplines each year 3) develop and extend the training of our graduate students to include substantial additional education in computational neuroscience and 4) expose a group of a dozen talented students primarily from other institutions to training and research in computational neuroscience. This last group of students will consist mostly of students from groups underrepresented in the field of computational neuroscience. The training will be broad and interdisciplinary including biological and psychological approaches on the experimental side and statistical, computational and mathematical on the quantitative side. [unreadable]